Green Chemistry

View Article Online

PAPER View Journal | View Issue

Microbial production of sebacic acid from a

Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.

Cite this: Green Chem., 2019, 21,

renewable source: production, purification, and
6491 polymerization†
Woo-Young Jeon,‡a Min-Jeong Jang,‡b Gyu-Yeon Park,‡a,b Hye-Jeong Lee,a
Sung-Hwa Seo,a Hee-Suk Lee,a Changpyo Han,a Heeun Kwon,a Ho-Chang Lee,c
Jong-Hwa Lee,c Yong-Taek Hwang,d Myung-Ock Lee,d Jeong-Gyu Lee,e
Hong-Weon Lee*a,b and Jung-Oh Ahn *a,b

Received 4th July 2019,
Accepted 31st October 2019
DOI: 10.1039/c9gc02274k
rsc.li/greenchem
Sebacic acid (SA) is an aliphatic ten-carbon dicarboxylic acid (1,10-decanedioic acid) with a variety of
industrial applications, including the production of plasticizers, lubricants, cosmetics, and plastics.
Currently, SA is produced exclusively from alkaline pyrolysis of castor oil. Herein, we present an environ-
mentally friendly green route of SA production from plant oil-derived sources by microbial ω-oxidation.
We genetically engineered β-oxidation-blocked diploid yeast Candida tropicalis, and created an effective
microbial cell factory with an increase of 46% in SA production by overexpression of genes involved in
ω-oxidation of hydrocarbons compared to the original strain. A biotransformation process of SA pro-
duction from decanoic acid methyl ester was developed to overcome the challenges of high-density cell
culture, substrate feed, substrate/intermediate toxicity, and foam generation. Fed-batch production of
engineered C. tropicalis resulted in a molar yield of above 98%, a productivity of 0.57 g L−1 h−1, and a final
titre of 98.3 g L−1 in a 5-litre fermenter and the results were successfully reproduced using a larger scale
50-litre fermenter. The produced SAs were successfully purified to >99.8% using acid precipitation and
recrystallization. Finally, bio-nylon 610 was successfully synthesized by polymerization of the purified SA
with hexamethylenediamine and showed thermal properties very similar to those of commercially avail-
able nylon 610. The processes developed and described in this study can be employed to produce and
isolate SA for the synthesis of bio-nylons, using environmentally friendly procedures based on microbial
biotransformation with potential industrial applications.

Introduction applied to dibasic polyesters to yield good flexibility, stability,

Sebacic acid (SA) is an aliphatic saturated 10-carbon dicar-
boxylic acid (DCA) with multiple applications.1 Its most
common application in the plastics industry is the production
of polyamide polymers (e.g. PA 6,10, PA 4,10, and PA 10,10)
where it confers important properties such as flexibility, hydro-
phobicity, durability, and low melting temperatures. SA is also
a
Biotechnology Process Engineering Centre, KRIBB, 30 Yeongudanji-ro, Ochang-eup,
Cheongwon-gu, Cheongju-si, 363-883, Korea. E-mail: hwlee@kribb.re.kr,
ahnjo@kribb.re.kr
b
Bioprocess Department, University of Science and Technology, 217 Gajeong-ro
Yuseong-gu, Daejeon, Korea
c
Aekyung Petrochemical Co., Ltd, 120, Sinseongnam-ro, Yuseong-gu, Daejeon, Korea
d
Lotte Chemical, 115 Gajeongbuk-ro, Yuseong-gu, Daejeon, 34110, Korea
e
SmallLab Co., Ltd, Techno Jungang-ro, Yuseong-gu, Deajeon, 243-15, Korea
† Electronic supplementary information (ESI) available: Data on impurity charac-
teristics. See DOI: 10.1039/c9gc02274k
‡ These authors equally contributed to this study.
chemical resistance, and solvent resistance.2 Moreover, SA is
used in the production of lubricants, such as dioctyl sebacate
(DOS) and dihexyl sebacate (DHS).3 Another interesting appli-
cation of SA is the development of chitosan- and collagen-
based 3D scaffold materials, which are biocompatible and
mechanically resistant.4
Most of the worldwide SA production occurs in China (over
20 000 metric tons are exported each year, more than 90% of
the global trade).5 It has been traditionally produced by chemi-
cal cleavage of ricinoleic acid or castor oil to SA and capryl
alcohol (2-octanol) under high temperature (about 280 °C) and
alkaline conditions (caustic pyrolysis).6 This process consumes
considerable amounts of H2SO4 and releases Na2SO4-contain-
ing solutions, which represent serious environmental risks.7
Additionally, Casda Biomaterials Co., Ltd, the world leader in
SA manufacture, announced increases in the price of SA due to
the strong decline in the price of 2-octanol, a by-product of SA
production.8 Nowadays, China’s new environmental policy has

This journal is © The Royal Society of Chemistry 2019 Green Chem., 2019, 21, 6491–6501 | 6491

Paper
been associated with factory shutdowns and arrested chemical
production of several compounds, including SA, owing to
environmental regulations.9
Microbial biotransformation and other biotechnological
approaches are gaining importance and have been increasingly
used to overcome the limitations and environmental issues
associated with chemical processes. Several natural metabolic
pathways have the potential to be explored for production of a
wide variety of molecules with several biotechnological appli-
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.
cations. Yeast, for example, can convert long-chain fatty acids
and n-alkanes into the corresponding DCAs via the ω-oxidation
pathway, whereby terminal (ω-position) carbons are oxidized
(Fig. 1).10,11 However, DCAs are naturally degraded by
β-oxidation in the peroxisome, where their aliphatic backbone
is continuously shortened via the cyclic cleavage of two carbon
acetyl-CoA moieties. The resulting acetyl-CoA molecules are
then metabolized through the tricarboxylic acid cycle, which
sustains growth or energy production.12 Using genetic engin-
eering techniques, one can block or stimulate a given pathway,
resulting in the accumulation of one or more desired metab-
olites. Microbial biotransformation of long-chain DCAs (C12–
C18) from n-alkane or fatty acids has been performed in
various alkane-assimilating yeasts like Candida tropicalis,
Yarrowia lipolytica, Torulopsis bombicola, Candida cloacae, and
Wickerhamiella sorbophila by blocking the β-oxidation
pathway.13–17 Moreover, non-alkane-assimilating
Saccharomyces cerevisiae was engineered for the production of
DCAs from fatty acids by a synthetic ω-oxidation pathway.18
Among bacteria, alkane-assimilating Pseudomonas aeruginosa
converted n-pentadecane to the corresponding DCA (C15).19
Fig. 1 Oxidation pathways for n-alkanes and fatty acids in yeast. The oxidation of n-alkane and fatty acids can occur by the ω-oxidation pathway of
the terminal methyl group to yield the α,ω-dicarboxylic acid. Fatty acids and α,ω-dicarboxylic acids are subjected to subsequent β-oxidation.
Deletion of the genes encoding acyl-CoA oxidase in β-oxidation is intended to solely oxidize n-alkane or fatty acids by ω-oxidation, thereby allowing
long-chain α,ω-dicarboxylic acid to accumulate.
6492 | Green Chem., 2019, 21, 6491–6501
View Article Online
Green Chemistry
Escherichia coli was engineered for the production of DCAs
(C12 and C14) by heterologous introduction of the ω-oxidation
pathway.20 However, their DCA titres were much lower (<1 g
L−1) than those of the engineered alkane-assimilating yeast
strains. Engineering of β-oxidation reversal in E. coli was
exploited to produce SA from glucose, but the titre was only
0.061 g L−1.21 A strain of S. cerevisiae, engineered by introdu-
cing a heterogeneous cytochrome P450, produced SA with a
very low titre of 0.012 g L−1.22
Candida tropicalis has been used as a sole strain for indus-
trial production of long-chain DCA because of its prominent
ability to oxidize alkanes and fatty acids due to strong
ω-oxidation activity, as reviewed by Huf et al. (2011) and Lee
et al. (2019).23,24 C. tropicalis have 10 cytochrome P450 genes,
(CYP52A12-20, CYP52D2) which have various substrate specifi-
cities and activities depending on the carbon length of the
substrates.1,25 This yeast species have been metabolically
engineered to transform long-chain fatty acids into DCAs by
blocking the β-oxidation pathway.13 The amplification of genes
encoding ω-hydroxylase complex components, which catalyse
the rate-limiting reaction in ω-oxidation pathway, enhances the
titre and productivity of DCA by C. tropicalis.26 In addition, the
production of DCA by microorganisms can be improved by the
optimization of parameters, such as medium composition,
culture conditions, and substrate feeding strategies.27–30
Recently, we investigated the toxicity of decanoic acid and
10-hydroxydecanoic acid on C. tropicalis. To improve the pro-
duction of SA, we used the continuous substrate feeding
method and the titre was 34.5 g L−1.31 These studies have
improved the production of long-chain DCAs with high titres
This journal is © The Royal Society of Chemistry 2019

Green Chemistry
and productivity. Recently, Cathay Biotechnology Inc.
(Shandong, China), leading microbial production of DCAs, is
currently producing C11-C18 long-chain DCAs from paraffin as
raw material, with an annual production of 20 000 metric
tons.32,33 Unlike other long-chain DCAs, relatively little
research has been carried out on the biological production of
SA.
Herein, we report an industrially applicable biotransform-
ation protocol using a genetically engineered C. tropicalis
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.
strain to produce SA from vegetable oil-derived substrates. Its
separation and purification process was also optimized, and
the purified SA molecules were polymerized for the synthesis
of bio-based nylon.
Materials and methods
Microorganisms
Escherichia coli DH5α [F− endA1 hsdR17(rK−mK−) supE44 thi-1
λ− recA1 gyrA96 ϕ80dlacZΔM15] was used as the host strain for
cloning and manipulation of plasmids. For the production of
SA, C. tropicalis strains used in this study are shown in Table 1.
These strains were genetically engineered from the β-oxidation-
blocked ATCC® 20962™ strain obtained from the American
Type Culture Collection (Manassas, VA, USA).
Amplification of genes involved in ω-oxidation pathway
The gene encoding orotidine 5-phosphate decarboxylase
(URA3) was amplified by PCR using as template the genomic
DNA of ATCC® 20336™ and primers URA-1 and URA-2, which
were designed based on the URA3 sequence (GenBank acces-
sion number AB006207). URA3 was inserted between the ApaI
and MluI sites of pGEM-T Easy Vector (Promega, Madison, WI,
USA) using infusion cloning kit (Takara, Kusatsu, Japan). The
resulting plasmid was termed as pGEM-URA3. For pop-out of
the URA3 selection marker by DNA recombination, two identi-
Table 1 Candida tropicalis strains used in this study
Strain Genotype
Ct0 (ATCC 20962) URA3/URA3 POX2/POX2 pox4/pox4 pox5/pox5
Ct1 ura3/ura3 POX2/POX2 pox4Δ::CYP52A18-URA3/pox4
pox5/pox5
Ct2 ura3/ura3 pox2Δ::CYP52A19-URA3/POX2 pox4/pox4
pox5/pox5
Ct3 ura3/ura3 POX2/POX2 pox4/pox4 pox5Δ::CPRB-
URA3/pox5
Ct4 ura3/ura3 pox2Δ::CYP52A19/POX2 pox4/pox4
pox5Δ::CPRB-URA3/pox5
Ct5 ura3/ura3 pox2Δ::CYP52A19/POX2 pox4Δ::
CYP52A18/pox4
pox5Δ::CPRB-URA3/pox5
Ct6 ura3/ura3 pox2Δ::CYP52A19/POX2 pox4/pox4Δ::
FAO1-URA3
pox5Δ::CPRB/pox5
Ct7 ura3/ura3 pox2Δ::CYP52A19/POX2 pox4Δ::
CYP52A18/pox4Δ::FAO1-URA3
pox5Δ::CPRB/pox5
This journal is © The Royal Society of Chemistry 2019
View Article Online
Paper
cal hisG DNA sequences from Bacillus subtilis were amplified
by PCR using two sets of primers SA-1/SA-2 and SA-3/SA-4.
HisG fragments were inserted between the BglII and EcoRI sites
on left side of URA3 and NotI and BamHI sites on right side of
URA3 in pGEM-URA3 respectively. The resulting plasmid was
termed as pGEM-HUH.
The promoter-CYP52A19-terminator was amplified by PCR
using genomic DNA of ATCC® 20336™ as template and
primers CYP-1 and CYP-2. The promoter-CYP52A19-terminator
fragment was inserted between the BglII site of pGEM-HUH
using infusion cloning kit (Takara, Kusatsu, Japan) and the
resulting plasmid was termed as pC19HUH. The promoter-
CYP52A18-terminator and promoter-CPRB-terminator were
amplified by PCR using the same genomic DNA as template
using primers CYP-3/CYP-4 and CPR-1/CPR-2 respectively. Each
fragment was inserted between the BglII site of pGEM-HUH in
the same manner, the resulting plasmids were termed as
pC18HUH and pCRHUH respectively. The promoter-FAO-ter-
minator was constructed by fusion PCR with TEF promoter and
FAO-terminator. TEF promoter and FAO-terminator were ampli-
fied by PCR using primers TEF-1/TEF-2 and FAO-1/FAO-2
respectively. The promoter-FAO-terminator was inserted
between the BglII site on pGEM-HUH, and the resulting
plasmid was termed as pFOHUH. The constructed plasmid
maps for each expression cassette are shown in the ESI Fig. S1.†
For the transformation of C. tropicalis an expression cas-
sette containing the target gene was inserted into the genome
site-specifically. The uracil auxotroph strain was constructed
by disrupting the URA3 selection marker gene of C. tropicalis
ATCC® 20962™ for the insertion of the cassette. Each
expression cassette containing the selectable marker gene was
inserted into the Candida genome through homologous
recombination with a sequence homologous to a portion of
POX2, POX4, and POX5 genes encoding acyl-CoA oxidases. The
URA+ transformants selected from the YNB plate were cultured
on the YNB-5FOA plate to pop-out the selectable marker gene.
The resulting URA3 auxotroph was transformed again by inser-
tion of an additional expression cassette. The primers used in
this study are shown in Table 2.
Preparation of C10 fatty acid methyl ester (FAME)
Decanoic acid derived from plant oil was purchased in bulk
from Emery Oleochemicals (Teluk Panglima Garang,
Malaysia). Decanoic acid (100 kg), 55.8 kg of methanol
(55.8 kg) and p-toluene sulfonic acid (1 kg) were injected into a
reactor and then refluxed for 7.5 h. After removal of residual
methanol in the reactor by evaporation, 55.8 kg of methanol
was injected into the reactor and reflux was performed for 6 h.
The same procedure was repeated with 28 kg of methanol and
1 hour of reflux. Then, 157 g of NaOH and 20 kg of distilled
water were added to the reactor, treated at 60 °C for 30 min,
and the lower NaOH solution was removed after layer separ-
ation. Distilled water (20 kg) was added to the reactor and the
lower layer of distilled water was removed through heat treat-
ment and layer separation in the same manner as before. Next,
the remaining crude C10 fatty acid methyl ester (FAME) was
Green Chem., 2019, 21, 6491–6501 | 6493
 View Article Online

Paper Green Chemistry

Table 2 Primers used in this study

Primers Sequences (5′ → 3′) Description

Ura-1 ccgcacctccgaattcgatctggtttggattgttggag URA3 cloning

Ura-2 accggtctggcggccgcgatcttgaagtcctcgtttgtg
SA-1 aattgggcccagatctcagaccggttcagacaggat pop-out vector cloning
SA-2 aaaccagatcgaattcggaggtgcggatatgaggta
SA-3 ttcaagatcgcggccgccagaccggttcagacaggat
SA-4 tccaacgcgtggatccggaggtgcggatatgaggta
CYP-1 aattgggcccagatctggccgacacactttcaacgg CYP52A19 cassette
CYP-2 aaccggtctgagatctcggtcacgttgttaggcatgc
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.

CYP-3 aattgggcccagatctttcaatgctgggccccgg CYP52A18 cassette

CYP-4 aaccggtctgagatctggggggattttgcatgcacc
CPR-1 aattgggcccagatctcccacccctcgcaggaat CPRb cassette
CPR-2 aaccggtctgagatctgccaatgccaatgccaaagaagt
TEF-1 aattgggcccagatcctgcagaaagaataacattagtaaataga TEF promoter
TEF-2 gccatatggattgattatttctttagatgtgtagag
FAO-1 gttatgctagctcgagctacaacttggccttggtcttcaag FAO1 cassette
FAO-2 gaaataatcaatccatatggctccatttttgcccgac
SA-5 gtcacgaccaccaacaaaag Cassettes insertion
SA-6 ccagaccgttgaaagtgtgtcggccccgaaacaaccataaagctgc
SA-7 atcgctacctcatatccgcacctccccaagttctccaaggacttg
SA-8 cgagtgcagaacggttcaac
SA-9 cctgaccctagatcatccatc
SA-10 aacaagtccggggcccagcattgaagtgttgatgacaaactcgtcag
SA-11 atcgctacctcatatccgcacctccgcaagaccgtcagaggttcc
SA-12 acaatcatgtcggattgttgg
SA-13 cgtacgagcctctgatattcttg
SA-14 gctctatttactaatgttattctttgcatacaactagcgagctacag
SA-15 atcgctacctcatatccgcacctcccactaacattgttcaaatcttcacg
SA-16 ctcgtcagattccttgtcaaatgtg
SA-17 cccaagaaagatccgagatcatc
SA-18 tcgaatgattcctgcgaggggtgggcaatctggcgtagacggagc
SA-19 atcgctacctcatatccgcacctccctccaacgtcagaatcccaag
SA-20 ccacaagcttgacgagcttc

collected and injected into a distillation apparatus for purifi-
cation. The distillation was carried out using a 30-stage distil-
lation column at a pressure of 10 to 20 torr, a reactor bottom
temperature of 149 to 164 °C, an upper reactor temperature of
110 to 138 °C, and a reflux ratio of 1. The purity of C10 FAME
made in this study was 99.45 ± 0.15% by GC analysis.
Cell culture
A single colony of each C. tropicalis strain grown on yeast
extract-peptone-dextrose (YPD) (2% glucose, 1% yeast extract,
2% Bacto-peptone, 2% agar) agar plate was inoculated into
20 mL of YPD liquid medium and incubated overnight at
30 °C on a rotary shaker (200 rpm). The culture was transferred
into an Erlenmeyer flask (1 L) containing 200 mL of YPD
medium, and incubated overnight at 30 °C on the rotary
shaker (200 rpm). After initial glucose was exhausted, 100 mL
of the culture was transferred into another Erlenmeyer flask
(1 L) with 100 mL of biotransformation medium (100 g L−1
glycerol, 13 g L−1 yeast nitrogen base without amino acids,
6 g L−1 yeast extract, 3.64 g L−1 KH2PO4, 3 g L−1 K2HPO4, and
1 g L−1 decane). After the total consumption of decane, 5 g L−1
of C10 FAME was added twice at 24-hour intervals.
Fermentation
A 5-litre fermenter system (CNS Inc., Daejeon, Korea) was
used for lab-scale biotransformation of C10 FAME to SA.
Inoculation was carried out by adding 10% (v/v) of the pre-
culture, which was performed in 2-litre Erlenmeyer flask
with 200 mL of YPD media. The fermenter culture con-
tained ( per litre) 90 g of glycerol, 1 g of MgSO4·7H2 O, 20 g
of yeast extract, 8 g of (NH4)2SO4, 2 g of KH2PO4, 0.1 g of
NaCl, 0.1 g of CaCl2·2H2O, and 1 mL of trace elements.
When initial glycerol was exhausted, a feeding medium con-
taining ( per litre) 700 g of glucose and 20 g of yeast extract
was continuously fed. For activation of ω-oxidation, decane
(Junsei Chemical, Kyoto, Japan) was added for 4 h after the
start of feeding media, and subsequently C10 FAME was
added. The temperature was controlled at 30 °C. The pH
was kept constant at 5.5 for cell growth phase and then it
was increased up to 7.8 for biotransformation phase with 6
N NaOH. Dissolved oxygen was maintained above 30% by
increasing the agitation speed, aeration and use of pure
oxygen.
This protocol was scaled up using a 50-litre fermenter
system (CNS Inc., Daejeon, Korea) for biotransformation of
C10 FAME to SA. Inoculation was carried out by adding 10%
(v/v) of the pre-culture, which was performed in four 5-litre
Erlenmeyer flasks with 500 mL of YPD media, in a final
volume of 2 L. Fermentation conditions in the 50-litre fermen-
ter, including the medium, temperature, pH, dissolved oxygen,
and procedures, were the same as in the 5-litre system
experiments.

6494 | Green Chem., 2019, 21, 6491–6501 This journal is © The Royal Society of Chemistry 2019

Green Chemistry
Analytical methods
The growth of yeast cells was monitored by measuring the
optical density at 600 nm using a UVICON930 spectrophoto-
meter (Kontron Analytical, Lucerne, Switzerland). Cell dry
weight was inferred from a predetermined conversion factor of
0.32 g per L per OD600. To determine residual glycerol or
glucose concentrations, 1 mL of culture broth was centrifuged,
and the concentration in the supernatant was measured using
a YSI 2900 model biochemistry analyser (YSI, Yellow Springs,
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.
OH, USA).
Decane, C10 FAME, decanoic acid, and SA were analysed
with a gas chromatograph (Master GC; Dani Instruments,
Cologno Monzese, Italy) equipped with an RTX-5 column
(Restek, Bellefonte, PA, USA). The Flame Ionization Detector
(FID) was used for GC analysis (model OPT 100 M – FID 86/C
Detection System). Oven temperature ranged from 70 to
237 °C, and the injector and detector temperatures were 280
and 300 °C, respectively. A 100 μL aliquot of 6 N H2SO4 was
added to 100 µL of culture broth. The sample was then mixed
with 300 µL of diethyl ether with tetradecanedioic acid, used
as an internal standard. Then, 100 μL of the upper organic
layer was collected and analysed by gas chromatography. For
decanoic acid and SA, this organic layer was silylated with the
same volume of N,O-bis-trimethylsilyl-trifluoroacetamide
before gas chromatography analysis.
Precipitation and purification of sebacic acid
The culture solution was centrifuged to separate the cells from
the medium. The pH of the supernatant was acidified to 4 by
the addition of HCl to precipitate SA. The precipitated SA was
separated from the supernatant by filtration and washed twice
using distilled water to remove soluble impurities. The crude
SA was mixed with decane at 150 °C and 500 rpm during 3 h
for recrystallization. The mixed solution was cooled at a rate of
1 °C min−1, followed by filtration to obtain SA. Recrystallized
SA was mixed with acetone for extraction and then filtered for
the removal of media-derived impurities (ESI Table S1†).
Immediately after extraction, the acetone solution containing
SA was decolorized by passing through activated carbon and
filtered using a rotary vacuum filter. Drying and granulation
were used to recover the purified SA.
Polymerization of nylon 610
For poly-condensation, the purified SA was mixed with hexa-
methylenediamine in water solution in a heated reaction
vessel under stirring. The reactor was heated to an inner temp-
erature of 230 °C and maintained at this temperature for 1 h
with an inner pressure of 17.2 bar. Then, the overpressure was
released within 1 h and the temperature was raised to 270 °C
during this step. Thereafter, the reaction mixture was main-
tained under a nitrogen stream for 80 min to support the post-
condensation reaction and to increase the molecular weight.
The polymer was then released through a die plate into water
for palletization. The melting temperature (Tm) and crystalliza-
tion temperature (Tc) of the produced nylon polymers were
This journal is © The Royal Society of Chemistry 2019
View Article Online
Paper
measured by differential scanning calorimetry (DSC, Q200, TA
Instruments, New Castle, DE, USA) according to ASTM D3418.
Then, the nylon 610 sample was melted gradually at 10 °C
min−1 and then cooled and then reheated at the same rate
from 30 to 250 °C. Degradation temperature (Td) of the pro-
duced nylon was measured thermogravimetric analysis (TGA,
Q2950, TA Instruments) according to ASTM E1131. Then, the
nylon 610 sample was heated from 40 to 900 °C at 20 °C min−1
under N2.
Statistics
Error bars represent the standard error of the mean from at
least three independent replicates. Minitab 17 (Minitab Inc,
State College, PA, USA) was used for statistical analyses of the
results. Paired t-tests were used to compare the production of
SA by differently genetically engineered strains; p values below
0.05 were considered to indicate significant differences.
Results and discussion
The overall processes for microbial production of SA from
renewable source, its purification and polymerization devel-
oped in this study are summarized in Fig. 2.
Strain engineering
In an effort to develop a microbial cell factory for industrial SA
production, we engineered the β-oxidation-blocked C. tropicalis
ATCC® 20962™ to overexpress genes involved in ω-oxidation
of n-alkane and fatty acids. Cytochrome P450 monooxygenase
(CYP450) and cytochrome P450 reductase complex (CPRb) con-
stitute the ω-hydroxylase enzyme complex, which catalyses the
rate-limiting first step of ω-oxidation pathway (Fig. 1).34
Accordingly, the coordinated amplification of CYP450 and
CPRb genes in C. tropicalis enhances the production of 1,12-
dodecanodioic acid or 1,14-tetradecodionic acid.35 In the
second step of ω-oxidation, fatty alcohol oxidase (FAO1) plays a
major role in the terminal oxidation of the alcohol group to an
aldehyde.36 Based on these observations, we attempted to
amplify the endogenous genes encoding the CYP52A18 and
CYP52A19 which are known to have effective oxidation activi-
ties on C10 substrates,35 CPRb, and FAO1 in C. tropicalis. The
target genes were site-specifically inserted into the genome by
URA3 split-marker recombination and bipartite cassettes
(Fig. 3a). As a result, several strains with various combinations
of gene amplifications were engineered (Fig. 3b and Table 1).
To evaluate the effect of the coordinated amplification of
genes involved in SA production, flask cultivation was per-
formed with the engineered C. tropicalis strains. Compared to
the parent strain (C. tropicalis 20962), individual over-
expression of CYP52A18, CYP52A19, and CPR increased SA
production by 15%, 9% and 6%, respectively (Fig. 3c). When
two or more of these genes were co-overexpressed, SA pro-
duction increased by 30% for CYP52A19 and CPR, and 32%
for CYP52A19, CYP52A18 and CPR compared to C. tropicalis
20962. These observations indicate that the combined over-
Green Chem., 2019, 21, 6491–6501 | 6495
 View Article Online

Paper Green Chemistry

Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.

Fig. 2 Schematic representation of microbial sebacic acid (SA) production, purification, and polymerization. Bio-polyamide 610 was produced by
polymerization of hexamethylenediamine with the purified SA.

Fig. 3 Engineering the β-oxidation-blocked Candida tropicalis ATCC® 20962™ by overexpressing endogenous genes involved in the ω-oxidation
pathway of n-alkanes and fatty acids. (a) Schematic representation of the expression cassettes and chromosome integration by homologous recom-
bination in C. tropicalis. (b) The chromosome integrated genes and integration locus in the engineered strains. (c) Comparison of DCA production
from C10 FAME in the engineered strains. The asterisks denote significant differences between the indicated groups as determined by paired t-tests
( p < 0.05).

6496 | Green Chem., 2019, 21, 6491–6501 This journal is © The Royal Society of Chemistry 2019

Green Chemistry
expression of monooxygenase and oxidoreductase was effective
in augmenting SA production, whereas expression of both
CYP52A18 and CYP52A19 with CPRb or FAO1 did not further
increase the SA yield significantly. Co-overexpression of
CYP52A19, CPR, and FAO1 (i.e. Ct6 strain) elicited an increase
of 46% in SA concentration compared to the parent strain
(Fig. 3c). Such an effect of FAO1 gene overexpression on the
production of SA stems from the accelerated conversion of
hydroxyl fatty alcohol to SA. Therefore, the strain Ct6, which
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.
had the highest SA productivity in flask culture, was used for
the biotransformation process with fermenters in the next
experiments.
Fermentation for sebacic acid production
Whole-cell biotransformation process using fatty acid deriva-
tives as a substrate is very challenging because of product/sub-
strate toxicity and foam generation issues, which limit pro-
ductivity and process stability. For these reasons, we have
developed a biotransformation process that uses high-density
Fig. 4 Microbial sebacic acid (SA) production process from C10 fatty acid methyl ester (FAME). (a) Schematic representation and (b) the three
phases of the microbial SA production process. (c, d) Kinetic profiles of microbial SA production processes in the (c) 5-litre fermenter system and (d)
50-litre fermenter system.
This journal is © The Royal Society of Chemistry 2019
View Article Online
Paper
cell culture and substrate feed, and minimizes substrate/inter-
mediate toxicity and foam generation. The biotransformation
process was divided into three culture phases: cell growth, acti-
vating ω-oxidation metabolism with alkane addition, and bio-
transformation from C10 FAME to SA (Fig. 4b). The conditions
at each step were first optimized using a 5-litre fermenter.
The initial phase for the biotransformation process is to
obtain cells at high density to act as biocatalysts. To do so,
glycerol and glucose were tested as initial carbon sources. At
60 g L−1 of initial glucose or glycerol, cells reached to the
density of 72.6 ± 5.4 OD600 for glucose and 103.4 ± 8.5 OD600
for glycerol. Furthermore, when glucose was used as a sub-
strate, organic acids and ethanol were produced, while gly-
cerol, a non-fermentable substrate, did not produce any of
these by-products (data not shown). Therefore, to achieve
high cell densities, around 150 OD600, 90 g L−1 of glycerol
was further used as initial carbon source. However, the
obstacle of glycerol utilization was the deficit of NADPH for-
mation because of minimal glycolytic flux from glyceralde-
Green Chem., 2019, 21, 6491–6501 | 6497

Paper
hyde-3-phosphate.37 Therefore, when the initial glycerol was
exhausted, glucose was fed at a low feed rate (3.0 g L−1 h−1) to
promote NADPH regeneration, a key factor for the production
of SA. This feeding rate maintained the carbon-limited state
in the fermenter.
The second phase was designed to induce the expression of
the enzymes involved in ω-oxidation. Alkanes are known to act
as universal inducers for activating the enzymes involved in
ω-oxidation. Among the alkanes, decane was chosen since its
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.
metabolism leads to the production of SA, other alkanes could
be converted to other unwanted DCAs as by-products. To
improve the SA production, decane induction was previously
compared between a single addition and continuous feeding
method.31 That study demonstrated that continuous feeding
was more effective than a single addition because it reduces
the accumulation of C10 hydroxy fatty acid (HFA), a toxic inter-
mediate to the cell. Continuous decane feeding was kept for
5 hours at 0.7 mL L−1 h−1, 60% of maximum decane uptake
rate (qdecane) determined by a single bolus addition. The pH
was gradually increased from 5.5 to 7.8, the optimal pH value
for DCA production.27 Glucose was continuously fed at the rate
of 3.0 g L−1 h−1 for cell maintenance and NADPH regeneration.
Decane was converted to the corresponding SA, and cell con-
centration was about 142 OD600 at the end of this phase.
The final phase was designed for the biotransformation of
C10 FAME into SA. The feed rate of the substrate is a crucial
parameter for optimization process due to its toxic effects on
the cells at high concentrations31 and, therefore, various sub-
strate feed rates were tested. The feed rates above 1.2 mL L−1
h−1 resulted in the accumulation of toxic intermediates, such
as decanoic acid and C10 HFA, which cause foaming and fer-
menter overflow. Therefore, C10 FAME feeding was started at
0.5 mL L−1 h−1, 70% of decane feeding rate in the second
phase (ω-oxidation activation). Then, the feed rate was
increased step-wise to 1.03 mL L−1 h−1 to enhance SA pro-
ductivity. Because the produced disodium SA generates foam
in the fermenter by reducing the surface tension, the air flow
rate was reduced to 0.1 vvm. The DO level (above 30%) was
maintained by the supply of pure O2 to 0.025 vvm to compen-
sate for the reduced air flow. The final SA concentration
reached 98.3 g L−1 with a volumetric productivity of 0.57 g L−1
h−1 and a molar yield above 98% (Fig. 4c). To our knowledge,
the titre, productivity, and process scale for SA production
described herein are the highest levels reported so far in the
world.
Next, the developed biotransformation process of SA pro-
duction was tested in a 50-litre system to test its feasibility for
industrial production. The biotransformation process in the
50-litre fermenter was performed following the same three
phases described above. The initial phase was carried out for
21 h and the maximum cell concentration was 162.1 OD600. To
minimize the generation of foams, the air flow rate was
reduced to 0.1 vvm and the DO level (above 30%) was main-
tained by the supply of pure O2 to 0.02 vvm. This 50-litre bio-
transformation process allowed a molar yield of above 98%, a
productivity of 0.68 g L−1 h−1, and a final titre of 90.9 g L−1
6498 | Green Chem., 2019, 21, 6491–6501
View Article Online
Green Chemistry
(Fig. 4d). It was confirmed that the scale up of the process for
SA production was feasible because the results in 50-litre stain-
less steel fermenter were similar to those in the 5-litre jar
fermenter.
Precipitation and purification of sebacic acid
To separate SA ( pKa = 4.72) from the other molecules in the
culture supernatant, we used acid precipitation by HCl
(Fig. 5a). In order to achieve the maximum precipitation yield
of SA, the pH was continuously lowered below its pKa, but
there was no difference in yield between pH 4 and lower pH
values. Therefore, acid precipitation was carried out at pH 4,
and the precipitated SA was filtered and washed with distilled
water. The precipitate was analysed, and the results indicated
that they consisted of 85.4% of SAs, 2.0% of other carbon-
chain (C11 and C12) diacids, 1.2% of monoacids (mainly C8
and C10), and 10.1% of impurities. Considering the purity
(99.45 ± 0.15%) of the substrate (C10FAME) used in this study,
other carbon-chain diacids may be formed from fatty acids in
the cell. Monoacids are formed due to the intermediates in the
ω-oxidation metabolism as observed in our previous work,31
but they must be removed because they terminate nylon
polymerization. The additional analysis of other impurities
indicated that they most likely originated from the media
because their components were similar to those found in yeast
extract (ESI Table S1†).
We developed a purification process that yields high-purity
SA that can be further polymerized. Moreover, this process
requires relatively less energy making it an economically feas-
ible process to develop. Noteworthy, it includes steps to
remove the monoacids that primarily terminate polymeriz-
ation, other impurities, and the yellowish colour (Fig. 5a). To
purify the crude SA precipitates separated from the culture
broth, the recrystallization process was first developed. The
proper choice of solvent is crucial in recrystallization and
therefore various solvents were tested in this study (data not
shown). Non-polar solvents showed higher yield than polar sol-
vents. This result is explained by the loss of residual SA dis-
solved in the polar solvent during the recrystallization process
due to the different solubility of SA in non-polar and polar sol-
vents. Thus, we chose non-polar decane to recrystallize SA,
which was able to remove monoacids completely and yielded
99.3% of SA recovery (Table 3). Secondly, solvent extraction
process was applied to remove media-derived impurities, and
acetone was used as a solvent due to its ability to solubilize SA
selectively and its high volatility. The acetone extraction
process successfully removed impurities (ESI Table S2†), and
SA was obtained at a purity of 96.4%. Finally, the decoloriza-
tion process was carried out with activated carbon, and the yel-
lowness index of the purified SAs was significantly reduced
from 206 ± 2 of the recrystallized SA to 9 ± 2 (Table 3). This
value was comparable to that of commercial SAs (10–15), and
the final purity and yield of SA were 98.0% (99.6% for diacids)
and 95.5%, respectively. It is too difficult to remove other
diacids due to their similar chemical properties to SA.
Furthermore, they do not terminate the polymerization, and
This journal is © The Royal Society of Chemistry 2019
 View Article Online

Green Chemistry Paper

Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.

Fig. 5 Separation and purification of sebacic acid (SA) from fermentation broths. (a) Process flow diagram and (b) photographs of the SA products.

Table 3 Summary of purification results of sebacic acid from fermentation broth

Sebacic Other Media-derived Other APHA

Purification step acid (g) diacids (g) Monoacid (g) impurities (g) impurities (g) Total weight (g) Yield (%) Purity (%) coloura

Crystallized sample 1.025.8 23.6 14.6 121.8 15.0 1200.8 — 85.4

Recrystallization 1018.3 31.1 0.0 121.8 6.8 1178.0 99.3 86.4 206 ± 2
from decane
Acetone extraction 1018.3 31.1 0.0 0.0 6.8 1056.2 99.3 96.4
Decolorization 979.5 17.7 0.0 0.0 2.7 999.9 95.5 98.0 9±2
a
APHA colour: a colour standard named for the American Public Health Association and defined by ASTM D1209.

This journal is © The Royal Society of Chemistry 2019 Green Chem., 2019, 21, 6491–6501 | 6499
 View Article Online

Paper Green Chemistry

therefore we did not develop additional process to remove

them in this study.
There have been few studies about the purification of other
long-chain diacids, and C12 diacid (1,12-dodecanedioic acid)
has been purified using the falling film crystallization method.
However, this process is not easy to apply industrially because
it requires great amounts of thermal energy and a sophisti-
cated temperature control system.38 For short-chain diacids,
many studies have reported the purification of succinic acid
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.

from microbial fermentation through various processes, such

as distillation-crystallization, precipitation, membrane separ-
ation, and solvent extraction.39 Although these purification
processes have achieved high purity (>99%), their yields
ranged from 60 to 93%. Consequently, this study, for the first
time, developed a process for the separation and purification
of SAs produced from microbial fermentation, and this
process shows the industrial applicability with 98% (99.6% for
diacids) of purity and 95.5% of yield at a low cost. Fig. 5b
shows the photographs of the SA product during the separ-
ation and purification steps of the SA.

Polymerization with nylon 610

To test whether the purified SA could be polymerized, we
attempted to synthesize nylon 610 by its polymerization with
hexamethylenediamine (NH2–(CH2)6–NH2). Nylon 610 was suc-
cessfully synthesized and compared to commercial nylon 610.
The chemical properties of the bio-nylon 610 synthesized in
this study were analysed using DSC and TGA (Fig. 6). Tm and
Tc were determined to be 223.5 °C and 188.1 °C, respectively,
Fig. 6 Analysis of the properties of bio-based nylon 610. (a) Differential
and Td, which indicates heat resistance, was determined to be scanning calorimetry (DSC) results. (b) Thermogravimetric analysis (TGA)
470.8 °C. These chemical properties indicated that bio-nylon results.
610 synthesized in this study is nearly equivalent compared to
commercial nylon 610 (Table 4). These thermal properties
suggest that the nylon 610 synthesized in this study can be Table 4 Comparison of material properties of bio-based nylon 610
used for coating of bristle like monofilaments of toothbrush synthesized in this study and commercially available chemical-based
and cables since it presented chemical resistance and excellent nylon 610
electric performance.
Conventional Bio-nylon 610
Mechanical property nylon 610 in this study

Melting temperature (Tm, °C) 222.4 223.5

Conclusions
In this study, a C. tropicalis strain was genetically engineered
to improve the SA production. Coordinate amplification of
monooxygenase, oxidoreductase, and fatty alcohol oxidase
showed the best results for the SA production in flask culture.
A biotransformation process was developed for the high-level
production of SA to overcome the challenges of high-density
cell culture, substrate feed, substrate/intermediate toxicity and
foam generation. Subsequently, we used this process in the
fed-batch production of engineered C. tropicalis produced
98.3 g L−1 of SA from renewable plant oil-derivatives with a
productivity of 0.57 g L−1 h−1 at 5-litre system scale. These
results were successfully reproduced using a larger 50-litre
bench scale system, which resulted in the highest value of SA
production so far to our knowledge. However, substrate toxicity
limited the SA productivity, and therefore the further develop-
Crystallization temperature (Tc, °C) 182.1 188.1
Degradation temperature (Td, °C) 470.4 470.8
ment of the strain, tolerant to substrate toxicity, will accelerate
the industrial application of this achievement. The separation
with acid precipitation and purification processes with solvent
extraction and recrystallization enhanced the purity of SA to
the polymer-grade. Bio-nylon 610 was successfully synthesized
by polymerizing hexamethylenediamine and SA. The syn-
thesized bio-nylon 610 showed properties similar to those of
commercially available nylon 610. Therefore, the microbial SA
biotransformation developed in this study has the potential to
replace the traditional caustic pyrolysis process to produce SA,
which is leading industries to shut down due to environmental
regulations.

6500 | Green Chem., 2019, 21, 6491–6501 This journal is © The Royal Society of Chemistry 2019

Green Chemistry
Author contributions
W. J, M. J and G. P designed and conducted the research, and
wrote the paper. H. L, S. S., H. L., C. H., H. K., H. L., J. L.,
Y. H., M. L., and J. L. helped to do the experiments and
analyse the samples; H. L. and J. A. conceived and supervised
the project. All authors wrote the manuscript together and
approved the final version.
Published on 01 November 2019. Downloaded by University of Toledo on 1/3/2020 4:19:25 AM.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
This study was supported by a grant from the KRIBB Research
Initiative Program, and industrial Strategic Technology
Development Program (10047873 & 20002734) funded by the
Ministry of Trade, Industry & Energy (MOTIE, Republic of Korea).
Notes and references
1 D. L. Craft, K. M. Madduri, M. Eshoo and C. R. Wilson,
Appl. Environ. Microbiol., 2003, 69, 5983–5991.
2 Polymerupdate, https://www.polymerupdate.com, accessed
October 2018.
3 W. A. Ahmed, J. Salimon and M. A. Yarmo, Int. J. Adv. Sci.
Eng. Inf. Technol., 2014, 4, 1–6.
4 G. Sailakshmi, T. Mitra and A. Gnanamani, Prog. Biomater.,
2013, 2, 11.
5 Wincom Inc., https://wincom-inc.com/products/sebacica-
cid, accessed October 2018.
6 A. K. Vasishtha, R. K. Trivedi and G. Das, J. Am. Oil Chem.
Soc., 1990, 67, 333–337.
7 L. Vaccaro, D. Lanari, A. Marrocchi and G. Strappaveccia,
Green Chem., 2014, 16, 3680–3704.
8 Arkema, http://www.arkema.com, accessed October 2018.
9 China’s new environmental policy is closing factories and
stopping chemical production, https://www.agchemigroup.
eu/cs/blog/prispevek/chinas-new-environmental-policy-
closing-factories-and-stopping-chemical-production, accessed
October 2018.
10 T. Sumita, T. Iida, A. Hirata, H. Horiuchi, M. Takagi and
A. Ohta, FEMS Microbiol. Lett., 2002, 214, 31–38.
11 J. F. Spencer, A. L. Spencer and C. Laluce, Appl. Microbiol.
Biotechnol., 2002, 58, 147–156.
12 E. H. Hettema and H. F. Tabak, Biochim. Biophys. Acta,
2000, 1486, 18–27.
13 S. Picataggio, K. Deanda and J. Mielenz, Mol. Cell. Biol.,
1991, 11, 4333–4339.
14 J. M. Nicaud, F. Thevenieau, M. T. Dall and R. Marchal,
WO2006064131A1, 2006.
This journal is © The Royal Society of Chemistry 2019
View Article Online
Paper
15 E. Wegner, US3796630A, 1974.
16 K. D. Green, M. K. Turner and J. M. Woodley, Enzyme
Microb. Technol., 2000, 27, 205–211.
17 H. Lee, C. Han, H. W. Lee, G. Park, W. Jeon, J. Ahn and
H. Lee, Biotechnol. Biofuels, 2018, 9, 310.
18 P. Buathong, N. Boonvitthya, G. Truan and
W. Chulalaksananukul, Int. Biodeterior. Biodegrad., 2019,
139, 70–77.
19 E. C. Chan, C. S. Cheng and Y. H. Hsu, J. Ferment. Bioeng.,
1997, 83, 157–160.
20 C. Sathesh-Pradu and S. Lee, J. Agric. Food Chem., 2015, 63,
8199–81208.
21 J. M. Clomburg, M. D. Blankschien, J. E. Vick, A. Chou,
S. Kim and R. Gonzalez, Metab. Eng., 2015, 28, 202–212.
22 L. Han, Y. Peng, Y. Zhang, W. Chen, Y. Lin and Q. Wang,
Front. Biol., 2017, 8, 2184.
23 S. Huf, S. Krugener, T. Hirth, S. Rupp and S. Zibek,
Eur. J. Lipid Sci. Technol., 2011, 113, 548–561.
24 H. Lee, Y. E. C. Sugiharto, H. Lee, W. Jeon, J. Ahn and
H. Lee, Appl. Microbiol. Biotechnol., 2019, 103,
1545–1555.
25 W. Lu, J. E. Ness, W. Xie, X. Zhang, J. Minshull and
R. A. Gross, J. Am. Chem. Soc., 2010, 132, 15451–15455.
26 S. Picataggio, T. Rohrer, K. Deanda, D. Lanning,
R. Reynolds, J. Mielenz and L. D. Eirich, Bio/Technology,
1992, 10, 894–898.
27 S. Liu, C. Li, X. Fang and Z. Cao, Enzyme Microb. Technol.,
2004, 34, 73–77.
28 S. Zibek, S. Huf, W. Wagner, T. Hirth and S. Rupp, Chem.
Ing. Tech., 2009, 81, 1797–1808.
29 J. Wang, J. Peng, H. Fan, X. Xiu, L. Xue, L. Wang, J. Su,
X. Yang and R. Wang, J. Ind. Microbiol. Biotechnol., 2018,
45, 971–981.
30 I. Funk, V. Sieber and J. Schmid, Sci. Rep., 2018, 7, 13842.
31 Y. E. C. Sugiharto, H. Lee, A. D. Fitriana, H. Lee, W. Jeon,
K. Park, J. Ahn and H. Lee, AMB Express, 2018, 8, 75.
32 K. Kroha, Int. News Fats, Oils Relat. Mater., 2004, 15, 568–
571.
33 P. Borg, G. Le, S. Lebrun and B. Pees, OCL, 2009, 16, 211–
214.
34 W. H. Eschenfeldt, Y. Zhang, H. Samaha, L. Stols,
L. D. Eirich, C. R. Wilson and M. I. Donnelly, Appl. Environ.
Microbiol., 2003, 69, 5992–5999.
35 T. Beardslee, S. Picataggio, E. D. Eirich and J. M. Laplaza,
WO2013006730, 2013.
36 Q. Cheng, D. Sanglard, S. Vanhanen, H. T. Liu, P. Bombelli,
A. Smith and A. R. Slabas, Biochim. Biophys. Acta, 2005,
1735, 192–203.
37 R. Yao, D. Xiong, H. Hu, M. Wakayama, W. Yu, X. Zhang
and K. Shimizu, Biotechnol. Biofuels, 2016, 9, 175.
38 L. Yu, L. Youzhi and Q. Xueqing, Chin. J. Chem. Eng., 2004,
12, 451–453.
39 K. K. Cheng, X. B. Zhao, J. Zeng and J. A. Zhang, Biofuels,
Bioprod. Biorefin., 2012, 6, 302–318.
Green Chem., 2019, 21, 6491–6501 | 6501